Radiation Segmentectomy: Are Lung Shunt Fractions Necessary for All Patients?

Nicolas Voutsinas, MD, recently presented an abstract at the Society for Interventional Radiology meeting in Austin, Texas. Dr. Voutsinas and his colleagues conducted a single-center, retrospective review of all radiation segmentectomy procedures performed with glass microspheres for the treatment of hepatocellular carcinoma, and their conclusions, although preliminary, suggest that lung shunt studies may not be necessary in certain well-selected patients. In this Q&A, Dr Voutsinas shares more information about the study.

Why did you decide to study this topic?

Radioembolization is a big area of interest in interventional radiology. We can now perform radiation segmentectomy as an effective treatment for hepatocellular carcinoma, but that means that patients will need to undergo a pre-planning procedure in order to assess anatomy and lung shunt fraction. Lung shunt fraction, if elevated, can result in radiation pneumonitis, which can be a devastating consequence to the patient. However, an additional procedure is always accompanied by concern about risk factors, so it benefits us to minimize extra procedures when possible. Our goal is to determine which patients are more likely to have low lung shunt fractions and which patients are predisposed to high lung shunt fractions so that we can better evaluate the patients.

Tell us about the design of the study.

We did a retrospective review from May 27, 2017 through February 28, 2018.  The reason we picked those dates was we wanted to be able to have a 6-month follow-up to assess for radiation pneumonitis, which usually presents 3 to 6 months after the procedure.

We looked at 88 patients at a single center and examined various clinical and imaging factors to see if any of them influenced the lung shunt fraction value. All the patients in the study had one solitary tumor and were receiving one segmentectomy treatment. The factors that we focused on were tumor size, AFP levels, and vascular invasion of the tumor.

Are there any other factors that you would ideally like to study?

We did look at several other factors, including age, demographics, and gender, but we did not elaborate on them in this abstract because there were no statistically significant results. We are going to expand on the data as we prepare to submit a manuscript for publication, and we will be able to look at additional factors more in depth.

What were your conclusions, and were your results surprising or as you expected?

In all of our patients (in an intention to treat analysis), the lung shunt fraction was below 12%. Our data showed that vascular invasion and elevated AFP levels correlated with statistically significant elevations in lung shunt fraction; however, tumor size did not. With regards to vascular invasion and elevated AFP levels, the fact that the lung shunt fraction was higher is not surprising as those factors are associated with more aggressive tumors. The size not having an effect was neither expected nor unexpected; it was simply an interesting finding.

Were there any challenges to conducting the study or processing it?

It was challenging to ensure we had a large enough sample size for conclusive results. For example, regarding the vascular invasion variable, 14 of the tumors showed vascular invasion, which means that 64 did not show invasion. Those are very different sample sizes, and that affects the statistical analysis. Our data are very preliminary, which is why we need to expand it before we submit it more formally.  

Can you tell us more about plans to expand the data?

We are going to include more years in the past in the analysis, as well as including the patients treated in the nearly a year that has passed since we submitted the initial abstract. We want to make sure that our results are not a sample error due to small size. We would, ideally, likely to have hundreds of patients to analyze.    

Looking even further ahead, if the signal is confirmed, what would be the next step?

Because the lung shunt fraction was low with all these patients in this specific population, these patients may not need to have this extra procedure. We might be able to potentially save carefully selected patients from undergoing a procedure and without risking complications. It’s very important to ensure this approach is safe though. We wouldn’t want to stop offering a procedure unless it was safe.  

Do you think the factors that your study reveals are applicable to the general population?

We hope so. That’s why we picked variables that are standard, such as size of tumor and AFP levels.  Additionally, vascular invasion is emphasized on pre-procedural imaging. For the purposes of the abstract and presentation, we wanted to keep the analysis simple and convey a few key points.

What is something that you want your colleagues to take away from your study at this point?

The important message is that we should be closely examining patient populations who are receiving Y-90 therapy, and make sure that each patient population is treated separately, and that we’re not treating a certain way just because everyone is treating that way. Instead, we should be giving directed and personalized therapy to the patients that are presenting to us.

Reference

N Voutsinas, A Lee, S Barazani, et al. Lung shunt fraction and radiation segmentectomy. Abstract 120. The Society for Interventional Radiology Meeting. March 24, 2019. Austin, Texas.